how can I apply Relu activation function in Levenberg_Marquardt algorithm for training neural network?

x = input';

t = target';

trainFcn = 'trainlm';

hiddenLayerSize = 10;

net = feedforwardnet(hiddenLayerSize,trainFcn);

net.divideParam.trainRatio = 70/100;

net.divideParam.valRatio = 15/100;

net.divideParam.testRatio = 15/100;

[net] = train(net,x,t);

Answer 1

[Varun Sai Alaparthi]

Hello Ashish,

You can use ‘poslin’ as transfer function to the hidden layer which is same as applying ‘ReLU’ activation function.

You can use this code for using 'poslin'

net.layers{1}.transferFcn = 'poslin'; 

I hope this information helps and please reach out for any further issues.